Background Little is known about the practice of ventilation management in patients with COVID-19. We aimed to describe the practice of ventilation management and to establish outcomes in invasively ventilated patients with COVID-19 in a single country during the first month of the outbreak.Methods PRoVENT-COVID is a national, multicentre, retrospective observational study done at 18 intensive care units (ICUs) in the Netherlands. Consecutive patients aged at least 18 years were eligible for participation if they had received invasive ventilation for COVID-19 at a participating ICU during the first month of the national outbreak in the Netherlands. The primary outcome was a combination of ventilator variables and parameters over the first 4 calendar days of ventilation: tidal volume, positive end-expiratory pressure (PEEP), respiratory system compliance, and driving pressure. Secondary outcomes included the use of adjunctive treatments for refractory hypoxaemia and ICU complications. Patient-centred outcomes were ventilator-free days at day 28, duration of ventilation, duration of ICU and hospital stay, and mortality. PRoVENT-COVID is registered at ClinicalTrials.gov (NCT04346342). FindingsBetween March 1 and April 1, 2020, 553 patients were included in the study. Median tidal volume was 6•3 mL/kg predicted bodyweight (IQR 5•7-7•1), PEEP was 14•0 cm H 2 O (IQR 11•0-15•0), and driving pressure was 14•0 cm H 2 O (11•2-16•0). Median respiratory system compliance was 31•9 mL/cm H 2 O (26•0-39•9). Of the adjunctive treatments for refractory hypoxaemia, prone positioning was most often used in the first 4 days of ventilation (283 [53%] of 530 patients). The median number of ventilator-free days at day 28 was 0 (IQR 0-15); 186 (35%) of 530 patients had died by day 28. Predictors of 28-day mortality were gender, age, tidal volume, respiratory system compliance, arterial pH, and heart rate on the first day of invasive ventilation. Interpretation In patients with COVID-19 who were invasively ventilated during the first month of the outbreak in the Netherlands, lung-protective ventilation with low tidal volume and low driving pressure was broadly applied and prone positioning was often used. The applied PEEP varied widely, despite an invariably low respiratory system compliance. The findings of this national study provide a basis for new hypotheses and sample size calculations for future trials of invasive ventilation for COVID-19. These data could also help in the interpretation of findings from other studies of ventilation practice and outcomes in invasively ventilated patients with COVID-19. Funding Amsterdam University Medical Centers, location Academic Medical Center.

No abstract

Background The intensity of ventilation, reflected by driving pressure (ΔP) and mechanical power (MP), has an association with outcome in invasively ventilated patients with or without acute respiratory distress syndrome (ARDS). It is uncertain if a similar association exists in coronavirus disease 2019 (COVID-19) patients with acute respiratory failure. Methods We aimed to investigate the impact of intensity of ventilation on patient outcome. The PRoVENT-COVID study is a national multicenter observational study in COVID-19 patients receiving invasive ventilation. Ventilator parameters were collected a fixed time points on the first calendar day of invasive ventilation. Mean dynamic ΔP and MP were calculated for individual patients at time points without evidence of spontaneous breathing. A Cox proportional hazard model, and a double stratification analysis adjusted for confounders were used to estimate the independent associations of ΔP and MP with outcome. The primary endpoint was 28-day mortality. Results In 825 patients included in this analysis, 28-day mortality was 27.5%. ΔP was not independently associated with mortality (HR 1.02 [95% confidence interval 0.88–1.18]; P = 0.750). MP, however, was independently associated with 28-day mortality (HR 1.17 [95% CI 1.01–1.36]; P = 0.031), and increasing quartiles of MP, stratified on comparable levels of ΔP, had higher risks of 28-day mortality (HR 1.15 [95% CI 1.01–1.30]; P = 0.028). Conclusions In this cohort of critically ill invasively ventilated COVID-19 patients with acute respiratory failure, we show an independent association of MP, but not ΔP with 28-day mortality. MP could serve as one prognostic biomarker in addition to ΔP in these patients. Efforts aiming at limiting both ΔP and MP could translate in a better outcome. Trial registration Clinicaltrials.gov (study identifier NCT04346342).

This is an open-access article distributed under the terms of the Creative Commons Attribution-Non Commercial-No Derivatives License 4.0 (CCBY-NC-ND), where it is permissible to download and share the work provided it is properly cited. The work cannot be changed in any way or used commercially without permission from the journal. OBJECTIVES:To analyze the available literature on the performance of artificial intelligence-generated clinical models for the prediction of serious life-threatening events in non-ICU adult patients and evaluate their potential clinical usage. DATA SOURCES:The PubMed database was searched for relevant articles in English literature from January 1, 2000, to January 23, 2022. Search terms, including artificial intelligence, machine learning, deep learning, and deterioration, were both controlled terms and free-text terms. STUDY SELECTION:We performed a systematic search reporting studies that showed performance of artificial intelligence-based models with outcome mortality and clinical deterioration. DATA EXTRACTION:Two review authors independently performed study selection and data extraction. Studies with the same outcome were grouped, namely mortality and various forms of deterioration (including ICU admission, adverse events, and cardiac arrests). Meta-analysis was planned in case sufficient data would be extracted from each study and no considerable heterogeneity between studies was present. DATA SYNTHESIS:In total, 45 articles were included for analysis, in which multiple methods of artificial intelligence were used. Twenty-four articles described models for the prediction of mortality and 21 for clinical deterioration. Due to heterogeneity of study characteristics (patient cohort, outcomes, and prediction models), meta-analysis could not be performed. The main reported measure of performance was the area under the receiver operating characteristic (AUROC) (n = 38), of which 33 (87%) had an AUROC greater than 0.8. The highest reported performance in a model predicting mortality had an AUROC of 0.935 and an area under the precision-recall curve of 0.96. CONCLUSIONS:Currently, a growing number of studies develop and analyzes artificial intelligence-based prediction models to predict critical illness and deterioration. We show that artificial intelligence-based prediction models have an overall good performance in predicting deterioration of patients. However, external validation of existing models and its performance in a clinical setting is highly recommended.